bupa1018's picture
Update app.py
8fde75c
raw
history blame
11.7 kB
import os
import json
import gradio as gr
import zipfile
import tempfile
import requests
import urllib.parse
import io
from huggingface_hub import HfApi, login
from PyPDF2 import PdfReader
from langchain_huggingface import HuggingFaceEmbeddings
from langchain_community.vectorstores import Chroma
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_groq import ChatGroq
from dotenv import load_dotenv
from langchain.docstore.document import Document
# Load environment variables from .env file
load_dotenv()
# Load configuration from JSON file
with open('config.json') as config_file:
config = json.load(config_file)
PERSIST_DIRECTORY = config["persist_directory"]
CHUNK_SIZE = config["chunk_size"]
CHUNK_OVERLAP = config["chunk_overlap"]
EMBEDDING_MODEL_NAME = config["embedding_model"]
LLM_MODEL_NAME = config["llm_model"]
LLM_TEMPERATURE = config["llm_temperature"]
GITLAB_API_URL = config["gitlab_api_url"]
HF_SPACE_NAME = config["hf_space_name"]
REPOSITORY_DIRECTORY = config["repository_directory"]
GROQ_API_KEY = os.environ["GROQ_API_KEY"]
HF_TOKEN = os.environ["HF_Token"]
login(HF_TOKEN)
api = HfApi()
def load_project_ids(json_file):
with open(json_file, 'r') as f:
data = json.load(f)
return data['project_ids']
def download_gitlab_repo():
print("Start the upload_gitRepository function")
project_ids = load_project_ids('repository_ids.json')
for project_id in project_ids:
print("Looping")
encoded_project_id = urllib.parse.quote_plus(project_id)
# Define the URL to download the repository archive
archive_url = f"{GITLAB_API_URL}/projects/{encoded_project_id}/repository/archive.zip"
# Download the repository archive
response = requests.get(archive_url)
archive_bytes = io.BytesIO(response.content)
# Retrieve the original file name from the response headers
content_disposition = response.headers.get('content-disposition')
if content_disposition:
filename = content_disposition.split('filename=')[-1].strip('\"')
else:
filename = 'archive.zip' # Fallback to a default name if not found
# Check if the file already exists in the repository
existing_files = api.list_repo_files(repo_id=HF_SPACE_NAME, repo_type='space')
target_path = f"{REPOSITORY_DIRECTORY}/{filename}"
print(f"Target Path: '{target_path}'")
print(f"Existing Files: {[repr(file) for file in existing_files]}")
if target_path in existing_files:
print(f"File '{target_path}' already exists in the repository. Skipping upload...")
else:
# Upload the ZIP file to the new folder in the Hugging Face space repository
print("Uploading File to directory:")
print(f"Archive Bytes: {repr(archive_bytes.getvalue())[:100]}") # Show a preview of bytes
print(f"Target Path in Repo: '{target_path}'")
api.upload_file(
path_or_fileobj=archive_bytes,
path_in_repo=target_path,
repo_id=HF_SPACE_NAME,
repo_type='space'
)
print("Upload complete")
def process_directory(directory):
all_texts = []
file_references = []
# if not os.path.exists(directory):
# raise ValueError(f"Directory {directory} does not exist.")
# Find all zip files in the directory
zip_files = [file for file in os.listdir(directory) if file.endswith('.zip')]
if not zip_files:
print("No zip files found in the directory.")
else:
for zip_filename in zip_files:
zip_file_path = os.path.join(directory, zip_filename)
# Create a temporary directory for each zip file
with tempfile.TemporaryDirectory() as tmpdirname:
# Unzip the file into the temporary directory
with zipfile.ZipFile(zip_file_path, 'r') as zip_ref:
zip_ref.extractall(tmpdirname)
# Process the temporary directory
temp_texts, temp_references = process_directory(tmpdirname)
all_texts.extend(temp_texts)
file_references.extend(temp_references)
for root, _, files in os.walk(directory):
for file in files:
file_path = os.path.join(root, file)
file_ext = os.path.splitext(file_path)[1]
if os.path.getsize(file_path) == 0:
print(f"Skipping an empty file: {file_path}")
continue
with open(file_path, 'rb') as f:
if file_ext in ['.rst', '.md', '.txt', '.html', '.json', '.yaml', '.py']:
text = f.read().decode('utf-8')
elif file_ext == '.pdf':
reader = PdfReader(f)
text = ""
for page in reader.pages:
text += page.extract_text()
elif file_ext in ['.svg']:
text = f"SVG file content from {file_path}"
elif file_ext in ['.png', '.ico']:
text = f"Image metadata from {file_path}"
else:
continue
all_texts.append(text)
file_references.append(file_path)
return all_texts, file_references
# Split text into chunks
def split_into_chunks(texts, references, chunk_size, chunk_overlap):
text_splitter = RecursiveCharacterTextSplitter(chunk_size=chunk_size, chunk_overlap=chunk_overlap)
chunks = []
for text, reference in zip(texts, references):
chunks.extend([Document(page_content=chunk, metadata={"source": reference}) for chunk in text_splitter.split_text(text)])
print(f"Total number of chunks: {len(chunks)}")
return chunks
# Setup Chroma
def setup_chroma(chunks, model_name="sentence-transformers/all-mpnet-base-v2", persist_directory="chroma_data"):
embedding_model = HuggingFaceEmbeddings(model_name=model_name)
vectorstore = Chroma.from_documents(chunks, embedding=embedding_model, persist_directory=persist_directory)
return vectorstore
# Setup LLM
def setup_llm(model_name, temperature, api_key):
llm = ChatGroq(model=model_name, temperature=temperature, api_key=api_key)
return llm
def query_chroma(vectorstore, query, k):
results = vectorstore.similarity_search(query, k=k)
chunks_with_references = [(result.page_content, result.metadata["source"]) for result in results]
# Print the chosen chunks and their sources to the console
print("\nChosen chunks and their sources for the query:")
for chunk, source in chunks_with_references:
print(f"Source: {source}\nChunk: {chunk}\n")
print("-" * 50)
return chunks_with_references
def rag_workflow(query):
docs = query_chroma(vectorstore, query, k=10)
context = "\n\n".join([doc for doc, _ in docs])
references = "\n".join([f"[{i+1}] {ref}" for i, (_, ref) in enumerate(docs)])
print(f"Context for the query:\n{context}\n")
print(f"References for the query:\n{references}\n")
prompt = f"You are an intelligent AI assistant who is very good in giving answers for anything asked or instructed by the user. Provide a clear and concise answer based only on the pieces of retrieved context. You must follow this very strictly, do not use anything else other than the retrieved context. If no related Information is found from the context, reply that you do not know. \n\nContext:\n{context}\n\nQuery: {query}"
response = llm.invoke(prompt)
return response.content, references
def initialize():
global vectorstore, chunks, llm
download_gitlab_repo()
all_texts, file_references = process_directory(REPOSITORY_DIRECTORY)
chunks = split_into_chunks(all_texts, file_references, CHUNK_SIZE, CHUNK_OVERLAP)
vectorstore = setup_chroma(chunks, EMBEDDING_MODEL_NAME, PERSIST_DIRECTORY)
llm = setup_llm(LLM_MODEL_NAME, LLM_TEMPERATURE, GROQ_API_KEY)
initialize()
# Gradio utils
def check_input_text(text):
if not text:
gr.Warning("Please input a question.")
raise TypeError
return True
def add_text(history, text):
history = history + [(text, None)]
yield history, ""
import gradio as gr
def bot_kadi(history):
user_query = history[-1][0]
response, references = rag_workflow(user_query)
history[-1] = (user_query, response)
# Format references for display with text passages
formatted_references = ""
docs = query_chroma(vectorstore, user_query, k=5)
for i, (doc, ref) in enumerate(docs):
formatted_references += f"""
<div style="border: 1px solid #ddd; padding: 10px; margin-bottom: 10px; border-radius: 5px;">
<h3 style="margin-top: 0;">Reference {i+1}</h3>
<p><strong>Source:</strong> {ref}</p>
<button onclick="var elem = document.getElementById('text-{i}'); var button = this; if (elem.style.display === 'block') {{ elem.style.display = 'none'; button.innerHTML = '&#9654; show source text'; }} else {{ elem.style.display = 'block'; button.innerHTML = '&#9660; hide source text'; }}">{{'&#9654; show source text'}}</button>
<div id="text-{i}" style="display: none;">
<p><strong>Text:</strong> {doc}</p>
</div>
</div>
"""
yield history, formatted_references
def main():
with gr.Blocks() as demo:
gr.Markdown("## Kadi4Mat - AI Chat-Bot")
gr.Markdown("AI assistant for Kadi4Mat based on RAG architecture powered by LLM")
with gr.Tab("Kadi4Mat - AI Assistant"):
with gr.Row():
with gr.Column(scale=10):
chatbot = gr.Chatbot([], elem_id="chatbot", label="Kadi Bot", bubble_full_width=False, show_copy_button=True)
user_txt = gr.Textbox(label="Question", placeholder="Type in your question and press Enter or click Submit")
with gr.Row():
with gr.Column(scale=1):
submit_btn = gr.Button("Submit", variant="primary")
with gr.Column(scale=1):
clear_btn = gr.Button("Clear", variant="stop")
gr.Examples(
examples=[
"Who is working on Kadi4Mat?",
"How do i install the Kadi-Apy library?",
"How do i install the Kadi-Apy library for development?",
"I need a method to upload a file to a record",
],
inputs=user_txt,
outputs=chatbot,
fn=add_text,
label="Try asking...",
cache_examples=False,
examples_per_page=3,
)
# with gr.Column(scale=3):
# with gr.Tab("References"):
# doc_citation = gr.HTML("<p>References used in answering the question will be displayed below.</p>")
user_txt.submit(check_input_text, user_txt, None).success(add_text, [chatbot, user_txt], [chatbot, user_txt]).then(bot_kadi, [chatbot], [chatbot])
submit_btn.click(check_input_text, user_txt, None).success(add_text, [chatbot, user_txt], [chatbot, user_txt]).then(bot_kadi, [chatbot], [chatbot])
clear_btn.click(lambda: None, None, chatbot, queue=False)
demo.launch()
if __name__ == "__main__":
main()